A new genus, Nothovernonia, from tropical Africa (Asteraceae or Compositae, Vernonieae)

Abstract Nothovernonia gen. nov., is described as a new genus for the tropical African Vernonia purpurea Sch.Bip. ex Walp. and Vernonia amblyolepis Bak, having cymiform inflorescences, pedunculate heads with differentiated foliiform bracts at the base, apiculate involucral bracts with scarious lateral margins, spiculiferous corolla lobes, and strongly spinose, sublopohate tricolporate pollen with the colpi meeting at the poles. The new genus belongs to the subtribe Centrapalinae and a key to the known DNA-sequenced genera of the subtribe is provided. The new species names are Nothovernonia purpurea (Sch.Bip. ex Walp.) H.Rob. and V.A.Funk, comb. nov., and Nothovernonia amblyolepis (Baker) H.Rob. & V.A.Funk, comb. nov.


Introduction
Many genera have been named in the tribe Vernonieae over the last 200 years, see Jones (1979bJones ( , 1981 and Jeffrey (1988) for citations of many of the names. During that time, some genera have been accepted when based on distinctions involving zygomorphic corollas (Elephantopus L, Pseudelephantopus Rohr or Stokesia L'Hér.), paleaceous receptacles (Lepidonia S.F. Blake, Centauropsis Boj. in DC.), syncephaly (Elephantopus Eremanthus Less., Lychnophora Mart.), a modified or reduced pappus (Elephantopus, Pseudelephantopus, Sparganophorus Vaill. ex Crantz, Pacourina Aubl., Centratherum Cass., Harleya S.F. Blake), or even sclerified tails on the anther when combined with strongly scandent habit and highly deciduous involucral bracts (Piptocarpha R. Br.). However, many other new genera whose members have some more general characters such as a capillary pappus, discrete heads, actinomorphic corollas, including many woody or arborecent types have not been accepted: Gymnanthemum Cass., Critoniopsis Sch. Bip., Strobocalyx (Blume ex DC.) Spach. Even some genera that were based on lophate pollen (Ambassa Steetz in Peters, Crystallopollen Steetz in Peters = Polydora Fenzl) have been summarily reduced to synonymy under the large, geographically improbable, core genus Vernonia Schreb. As a result, Vernonia s.l. was traditionally defined by what it "was not" rather than by what it "was" (Jones 1977, 1979b, 1981, Jeffrey 1988. Detailed morphological work and molecular phylogenics have finally forced the dismemberment of that unnatural genus, and true Vernonia s.s. is now known to be part of native floras only in the Americas (Robinson 1999a, b). As a result, the many paleotropical species that have been placed in the genus Vernonia are gradually being shifted into other genera (Robinson 1999a, 2007, Keeley and Robinson 2009. Although much progress has been made, there are still unplaced species, some of which can be assigned to new genera with reasonable assurance even though the defining characters are not necessarily macroscopic. One such genus is described here for a pair of species in tropical Africa including the common Vernonia purpurea Sch.Bip. ex Walp. The species known as Vernonia purpurea is a coarse herb ( Fig. 1) placed by Jeffrey (1988) in his Vernonia Subgroup A, a subgroup that included the genera Centrapalus Cass. and Vernonella Sond. as synonyms. More recent studies have shown that both Centrapalus and Vernonella are valid genera that apparently belong to different subtribes, Centrapalus to the Centrapalinae and Vernonella to the Linziinae (Robinson and Skvarla 2010). The same study has shown that Vernonia purpurea does not belong to either of the latter genera, differing in corolla color, corolla lobe pubescence, sweeping hairs of the style, and pollen structure. Isawumi (2008) placed V. purpurea in the genus Linzia Sch.Bip. ex Walp., but V. purpurea has neither the denticulations on the lateral margins of the involucral bracts, the bluish corollas, nor the lophate pollen with mural spurs projecting into the colpi that are characteristic of Linzia.
The goal of this study is to recognize the new genus and provide the proper nomenclature for a larger study on the subfamily Cichoroideae.

Methods
The morphology was studied using herbarium material, most of which was from the U.S. National Herbarium in Washington, D.C. Microscopic characters were examined via plant material mounted on microscope slides in semi-permanent, water misable Hoyer's Solution (Anderson 1954). Pollen grains were mostly treated with acetolysis  (Erdtman 1960), followed by staining with Osmium-thio-carbohydrazide solutions and sputter coating with gold/palladium (Robinson and Skvarla 2006, 2007, Robinson et al. 2008. Unacetylized grains were rehydrated in water or alcohol directly from herbarium sheets and similarly sputter coated. Observations were made with a JEOL 880 (Samuel Roberts Microscopy Laboratory, University of Oklahoma), or a LEICA 440 and AMRAY 1810 (National Museum of Natural History, Washington DC) scanning electron microscopes (SEM), all equipped with lanthanum hexaborate (LaB 6 ) electron sources.
Scanned images of the syntypes of Vernonia amblyolepis Baker were sent by the Herbarium at the Royal Botanic Gardens, Kew.
The position on the molecular phylogeny was determined by analyzing DNA sequence data of ITS, ndhF, trnL-F and matK. The details of the molecular work will be published as part of a larger study of the subfamily Cichorioideae (Funk and Chan 2009, unpublished data, Funk pers comm) (Fig. 2). Before the subfamily paper can be submitted, some nomenclatural issues must be straightened out, hence the necessity of this paper.

Discussion
The new genus, named here as Nothovernonia, can be distinguished from Centrapalus by the erect rather than decumbent bases of the stems, the apiculate laterally scarious rather than long herbaceous non-scarious tips on the involucral bracts, and the purple or lilac rather than blue-purple corollas (Fig. 3, B-D). The corolla color may be most useful in the field, with Centrapalus usually being bluish and Nothvernonia apparently never being blue. The stout spreading sweeping hairs of the style branches are different from the slender usually more appressed hairs in Centrapalus. The achenes are similar in having narrow though small raphids, many idioblasts on the surface, setulae with pairs of cells usually separated to a third or more of their length, but differ in the cells of the setulae being fused in the basal one to two thirds and the pappus being white to rufous. In Centrapalus the cells of the setulae are separate and essentially solitary to near the base, and the pappus is sordid grayish. Robinson  Pope. The two species treated here in Nothovernonia have pollen that is similar to but not identical with Centrapalus. They have long spines arranged in a sublophate pattern that is much less irregular than Cenrapalus and the colpi are not truncated, rather they reach to and meet at the poles (Robinson and Skvarla 2010) This type of pollen is widely distributed in the tribe and of uncertain use at this point. Never-the-less, pollen is frequently very useful in the tribe and it is essential to document the characters of the pollen for future comparative studies.
The two species treated here have their corolla color and stem bases more like those of Vernonella, but the colpi of the pollen are not truncated and thus failing to reach the poles, the involucral bracts are not scarious across the tips, the corolla lobes are spiculiferous with stiff hairs outside distally, and the achenes have setulae with cell pairs separated nearly half way but not to the base. In addition, the two species differ from both Centrapalus and Vernonella by the small to large differentiated foliiform bracteoles at the base of the head. This contrasts with the loose but undiffentiated narrow involucral bracts at the base of the head in Centrapalus and the small and broad but otherwise undifferentated bracts at the base of the head in Vernonella.
On the basis of structural features, the genus is clearly distinct, but the position nearer Centrapalus of the subtribe Centrapalinae or closer to Vernonella of the Linziinae, has been resolved with certainty only with DNA sequencing. Sequence data are available for Centrapalus, Parapolydora and Nothovernonia purpurea (see below). Members of Vernonella have not been sequenced, but the subtribal placement is derived from the report of elemanolide sesquiterpene lactones from one of the species, V. praemorsa (Muschl.) H.Rob. & Skvarla (Jakupovic et al. 1987), a type of sesquiterpene lactone known in the Vernonieae almost exclusively from the more basal members of the tribe. Figure 2 is the result of the analysis of DNA sequence data and shows all of the subtribes in the Vernonieae and all the known genera within the Centrapalinae. Within this scheme, Vernonia purpurea belongs in the broad group first placed by Robinson (1999a) in the Erlangeinae, but subsequently shown by DNA sequencing to belong to the distinct subgroup seen in Keeley et al. (2007)  This group is now recognized as part of the subtribe Centrapalinae H.Rob. (Keeley and Robinson 2009). Of the more restricted clade containing the closest relatives of Centrapalus and Vernonia purpurea, according to the DNA sequence results, Cabobanthus is the most basal in the group; it differs from the more highly nested genera by its lophate pollen. Parapolydora, the closest to Centraplus, shares the characters of numerous idioblasts on the achenes and the deeply divided or mostly solitary elongate cells of the setulae on the achenes. Nothovernonia, the genus named here based on Vernonia purpurea, falls somewhat outside the Centrapalus/Parapolydora group with strongly acuminate tips on the involucral bracts, foliiform bracts subtending the heads, and achenes with short narrow raphids, and more extensively fused elongate cells in the setulae. It is notable, however that the three closely related genera, Nothovernonia, Centrapalus and Parapolydora all share setulae with cells separated to at least one third of their length, more separated than in the less closely related genera of the Centrapalinae.

taxonomic treatment
Genera known to be in the subtribe Centrapalinae based on DNA sequence data are shown in Figure 2.
Key to the Genera known to be in the Centrapalinae (for a key to the subtribes of Vernonieae see Keeley and Robinson 2009) Figure 2. A phylogeny of the subtribes of the Vernonieae, with a detailed look at the subtribe Centropalineae. This tree is the strict consensus tree from a PAUP analysis that produced 28 trees. The data included were ITS, ndhF, trnL-F, and matK. The analysis included all known sequenced genera of the Centropalineae and is part of a larger analysis of the subfamily Cichorioideae.

tricolporatis differt. A Centrapalus in basis erectis et bracteis involcri apiculatis lateraliter scariosis differt.
Type. Vernonia purpurea Sch.Bip. ex Walp. Description. Coarse branching herbs to 0.7 m tall, stems erect from base, distinctly ribbed, pilose with spreading simple hairs. Leaves alternate with petioles 0.2-2.0 mm long; leaf blades oblong to lanceolate, sparsely pilose above with prominent persistent bases on the hairs, lower surface coarsely and densely pilose on major veins, surface with many glandular dots, secondary veins pinnate, ca. 6 pairs. Inflorescences terminal and from axils of reduced upper leaves, distinctly cymiform with distinct short to long peduncles; heads broadly campanulate, with minute to large foliose bracts at base; involucral bracts in ca. 5 series, strongly gradate, appressed, ovate to narrowly oblong, with dark median stripe, apex apiculate with distinct dark or rarely pale awn, with pale scarious lateral margins, outer surface with numerous arachnoid hairs from median band, spreading tranversely as the head expands; receptacles flat, without pales or hairs. Florets ca. 30-65 per head; corollas purple, with lobes and upper throat exceeding the pappus at anthesis, glanduliferous on throat, lobes spiculiferous with stout straight hairs distally on outer surface; anther thecae with small sterile border at base, endothecial cells oblong with sinuous lateral walls, apical appendage triangular, firm. Style base with annuliform sclerified node, distally with stout spreading sweeping hairs covering backs of style branches and upper 1 mm of style shaft. Achenes prismatic, 8-10-ribbed, with numerous idioblasts on surface, with small narrow raphids, setulae with pairs of cells fused together beyond basal 1/3; pappus with inner series of many crowded capillary bristles, white or rufous, less than 2/3 as long as corollas; outer pappus series of shorter, crowded lanceolate scales.
Chromosome number n = 9 (Jones 1979a   Unrecognized taxa. Jeffrey (1988) treated V. kandtii as a separate species that included V. pratensis Hiern and V. exasperata H. Wild in its synonymy. Vernonia kandtii is separated by Jeffrey (1988), with some doubt, by his couplet 106 as follows: Inflorescence copious, with a number of branches arising from the uppermost leaf-axils; ultimate peduncles short, so that the capitula appear to be in clusters .. The specimens from Malawi and Uganda that were examined for this study show variation in the structure of the inflorescence and therefore, do not support a separate species status for V. kandtii.